Note

Ray 2.40 uses RLlib’s new API stack by default. The Ray team has mostly completed transitioning algorithms, example scripts, and documentation to the new code base.

If you’re still using the old API stack, see New API stack migration guide for details on how to migrate.

Ray RLlib API

Tip

We’d love to hear your feedback on using RLlib - sign up to our forum and start asking questions!

This section contains an overview of RLlib’s package- and API reference. If you think there is anything missing, please open an issue on Github.

• Algorithm Configuration API
  • Constructor
  • Builder methods
  • Properties
  • Getter methods
  • Public methods
  • Configuration methods
• Algorithms
  • Building Custom Algorithm Classes
  • Algorithm API
• Callback APIs
  • Methods to implement for custom behavior
  • RLlibCallback
  • Callbacks invoked in Algorithm
  • Callbacks invoked in EnvRunner
• Environments
  • Env Vectorization
  • External Envs
  • Environment API Reference
• RLModule APIs
  • RLModule specifications and configurations
  • RLModule API
  • MultiRLModule API
• Distribution API
  • Base Distribution class
• LearnerGroup API
  • Configuring a LearnerGroup and Learner actors
  • Constructing a LearnerGroup
• Learner API
  • Constructing a Learner
  • Implementing a custom RLModule to fit a Learner
  • Performing updates
  • Computing losses
  • Configuring optimizers
  • Gradient computation
  • Saving, loading, checkpointing, and restoring states
  • Adding and removing modules
• Offline RL API
  • Configuring Offline RL
  • Configuring Offline Recording EnvRunners
  • Constructing a Recording EnvRunner
  • Constructing OfflineData
  • Sampling from Offline Data
  • Constructing an OfflinePreLearner
  • Transforming Data with an OfflinePreLearner
• Replay Buffer API
  • Replay Buffer Base Classes
  • Public Methods
  • Multi Agent Buffers
  • Utility Methods
• RLlib Utilities
  • Scheduler API
  • Framework Utilities
  • Checkpoint utilities